Examples with BenchmarkSchemaInfo io.druid.benchmark.datagen.BenchmarkSchemaInfo used on opensource projects

Example 1 with BenchmarkSchemaInfo

use of io.druid.benchmark.datagen.BenchmarkSchemaInfo in project druid by druid-io.

the class GroupByTypeInterfaceBenchmark method setupQueries.

private void setupQueries() {
    // queries for the basic schema
    Map<String, GroupByQuery> basicQueries = new LinkedHashMap<>();
    BenchmarkSchemaInfo basicSchema = BenchmarkSchemas.SCHEMA_MAP.get("basic");
    {
        // basic.A
        QuerySegmentSpec intervalSpec = new MultipleIntervalSegmentSpec(Arrays.asList(basicSchema.getDataInterval()));
        List<AggregatorFactory> queryAggs = new ArrayList<>();
        queryAggs.add(new LongSumAggregatorFactory("sumLongSequential", "sumLongSequential"));
        GroupByQuery queryString = GroupByQuery.builder().setDataSource("blah").setQuerySegmentSpec(intervalSpec).setDimensions(Lists.<DimensionSpec>newArrayList(new DefaultDimensionSpec("dimSequential", null))).setAggregatorSpecs(queryAggs).setGranularity(Granularity.fromString(queryGranularity)).build();
        GroupByQuery queryLongFloat = GroupByQuery.builder().setDataSource("blah").setQuerySegmentSpec(intervalSpec).setDimensions(Lists.<DimensionSpec>newArrayList(new DefaultDimensionSpec("metLongUniform", null), new DefaultDimensionSpec("metFloatNormal", null))).setAggregatorSpecs(queryAggs).setGranularity(Granularity.fromString(queryGranularity)).build();
        GroupByQuery queryLong = GroupByQuery.builder().setDataSource("blah").setQuerySegmentSpec(intervalSpec).setDimensions(Lists.<DimensionSpec>newArrayList(new DefaultDimensionSpec("metLongUniform", null))).setAggregatorSpecs(queryAggs).setGranularity(Granularity.fromString(queryGranularity)).build();
        GroupByQuery queryFloat = GroupByQuery.builder().setDataSource("blah").setQuerySegmentSpec(intervalSpec).setDimensions(Lists.<DimensionSpec>newArrayList(new DefaultDimensionSpec("metFloatNormal", null))).setAggregatorSpecs(queryAggs).setGranularity(Granularity.fromString(queryGranularity)).build();
        basicQueries.put("string", queryString);
        basicQueries.put("longFloat", queryLongFloat);
        basicQueries.put("long", queryLong);
        basicQueries.put("float", queryFloat);
    }
    {
        // basic.nested
        QuerySegmentSpec intervalSpec = new MultipleIntervalSegmentSpec(Arrays.asList(basicSchema.getDataInterval()));
        List<AggregatorFactory> queryAggs = new ArrayList<>();
        queryAggs.add(new LongSumAggregatorFactory("sumLongSequential", "sumLongSequential"));
        GroupByQuery subqueryA = GroupByQuery.builder().setDataSource("blah").setQuerySegmentSpec(intervalSpec).setDimensions(Lists.<DimensionSpec>newArrayList(new DefaultDimensionSpec("dimSequential", null), new DefaultDimensionSpec("dimZipf", null))).setAggregatorSpecs(queryAggs).setGranularity(Granularities.DAY).build();
        GroupByQuery queryA = GroupByQuery.builder().setDataSource(subqueryA).setQuerySegmentSpec(intervalSpec).setDimensions(Lists.<DimensionSpec>newArrayList(new DefaultDimensionSpec("dimSequential", null))).setAggregatorSpecs(queryAggs).setGranularity(Granularities.WEEK).build();
        basicQueries.put("nested", queryA);
    }
    SCHEMA_QUERY_MAP.put("basic", basicQueries);
}

Example 2 with BenchmarkSchemaInfo

use of io.druid.benchmark.datagen.BenchmarkSchemaInfo in project druid by druid-io.

the class SelectBenchmark method setupQueries.

private void setupQueries() {
    // queries for the basic schema
    Map<String, Druids.SelectQueryBuilder> basicQueries = new LinkedHashMap<>();
    BenchmarkSchemaInfo basicSchema = BenchmarkSchemas.SCHEMA_MAP.get("basic");
    {
        // basic.A
        QuerySegmentSpec intervalSpec = new MultipleIntervalSegmentSpec(Arrays.asList(basicSchema.getDataInterval()));
        Druids.SelectQueryBuilder queryBuilderA = Druids.newSelectQueryBuilder().dataSource(new TableDataSource("blah")).dimensionSpecs(DefaultDimensionSpec.toSpec(Arrays.<String>asList())).metrics(Arrays.<String>asList()).intervals(intervalSpec).granularity(Granularities.ALL).descending(false);
        basicQueries.put("A", queryBuilderA);
    }
    SCHEMA_QUERY_MAP.put("basic", basicQueries);
}

Example 3 with BenchmarkSchemaInfo

use of io.druid.benchmark.datagen.BenchmarkSchemaInfo in project druid by druid-io.

the class SqlBenchmark method setup.

@Setup(Level.Trial)
public void setup() throws Exception {
    tmpDir = Files.createTempDir();
    log.info("Starting benchmark setup using tmpDir[%s], rows[%,d].", tmpDir, rowsPerSegment);
    if (ComplexMetrics.getSerdeForType("hyperUnique") == null) {
        ComplexMetrics.registerSerde("hyperUnique", new HyperUniquesSerde(HyperLogLogHash.getDefault()));
    }
    final BenchmarkSchemaInfo schemaInfo = BenchmarkSchemas.SCHEMA_MAP.get("basic");
    final BenchmarkDataGenerator dataGenerator = new BenchmarkDataGenerator(schemaInfo.getColumnSchemas(), RNG_SEED + 1, schemaInfo.getDataInterval(), rowsPerSegment);
    final List<InputRow> rows = Lists.newArrayList();
    for (int i = 0; i < rowsPerSegment; i++) {
        final InputRow row = dataGenerator.nextRow();
        if (i % 20000 == 0) {
            log.info("%,d/%,d rows generated.", i, rowsPerSegment);
        }
        rows.add(row);
    }
    log.info("%,d/%,d rows generated.", rows.size(), rowsPerSegment);
    final PlannerConfig plannerConfig = new PlannerConfig();
    final QueryRunnerFactoryConglomerate conglomerate = CalciteTests.queryRunnerFactoryConglomerate();
    final QueryableIndex index = IndexBuilder.create().tmpDir(new File(tmpDir, "1")).indexMerger(TestHelper.getTestIndexMergerV9()).rows(rows).buildMMappedIndex();
    this.walker = new SpecificSegmentsQuerySegmentWalker(conglomerate).add(DataSegment.builder().dataSource("foo").interval(index.getDataInterval()).version("1").shardSpec(new LinearShardSpec(0)).build(), index);
    final Map<String, Table> tableMap = ImmutableMap.<String, Table>of("foo", new DruidTable(new TableDataSource("foo"), RowSignature.builder().add("__time", ValueType.LONG).add("dimSequential", ValueType.STRING).add("dimZipf", ValueType.STRING).add("dimUniform", ValueType.STRING).build()));
    final Schema druidSchema = new AbstractSchema() {

        @Override
        protected Map<String, Table> getTableMap() {
            return tableMap;
        }
    };
    plannerFactory = new PlannerFactory(Calcites.createRootSchema(druidSchema), walker, CalciteTests.createOperatorTable(), plannerConfig);
    groupByQuery = GroupByQuery.builder().setDataSource("foo").setInterval(new Interval(JodaUtils.MIN_INSTANT, JodaUtils.MAX_INSTANT)).setDimensions(Arrays.<DimensionSpec>asList(new DefaultDimensionSpec("dimZipf", "d0"), new DefaultDimensionSpec("dimSequential", "d1"))).setAggregatorSpecs(Arrays.<AggregatorFactory>asList(new CountAggregatorFactory("c"))).setGranularity(Granularities.ALL).build();
    sqlQuery = "SELECT\n" + "  dimZipf AS d0," + "  dimSequential AS d1,\n" + "  COUNT(*) AS c\n" + "FROM druid.foo\n" + "GROUP BY dimZipf, dimSequential";
}

Example 4 with BenchmarkSchemaInfo

use of io.druid.benchmark.datagen.BenchmarkSchemaInfo in project druid by druid-io.

the class TimeseriesBenchmark method setupQueries.

private void setupQueries() {
    // queries for the basic schema
    Map<String, TimeseriesQuery> basicQueries = new LinkedHashMap<>();
    BenchmarkSchemaInfo basicSchema = BenchmarkSchemas.SCHEMA_MAP.get("basic");
    {
        // basic.A
        QuerySegmentSpec intervalSpec = new MultipleIntervalSegmentSpec(Arrays.asList(basicSchema.getDataInterval()));
        List<AggregatorFactory> queryAggs = new ArrayList<>();
        queryAggs.add(new LongSumAggregatorFactory("sumLongSequential", "sumLongSequential"));
        queryAggs.add(new LongMaxAggregatorFactory("maxLongUniform", "maxLongUniform"));
        queryAggs.add(new DoubleSumAggregatorFactory("sumFloatNormal", "sumFloatNormal"));
        queryAggs.add(new DoubleMinAggregatorFactory("minFloatZipf", "minFloatZipf"));
        queryAggs.add(new HyperUniquesAggregatorFactory("hyperUniquesMet", "hyper"));
        TimeseriesQuery queryA = Druids.newTimeseriesQueryBuilder().dataSource("blah").granularity(Granularities.ALL).intervals(intervalSpec).aggregators(queryAggs).descending(false).build();
        basicQueries.put("A", queryA);
    }
    {
        QuerySegmentSpec intervalSpec = new MultipleIntervalSegmentSpec(Arrays.asList(basicSchema.getDataInterval()));
        List<AggregatorFactory> queryAggs = new ArrayList<>();
        LongSumAggregatorFactory lsaf = new LongSumAggregatorFactory("sumLongSequential", "sumLongSequential");
        BoundDimFilter timeFilter = new BoundDimFilter(Column.TIME_COLUMN_NAME, "200000", "300000", false, false, null, null, StringComparators.NUMERIC);
        queryAggs.add(new FilteredAggregatorFactory(lsaf, timeFilter));
        TimeseriesQuery timeFilterQuery = Druids.newTimeseriesQueryBuilder().dataSource("blah").granularity(Granularities.ALL).intervals(intervalSpec).aggregators(queryAggs).descending(false).build();
        basicQueries.put("timeFilterNumeric", timeFilterQuery);
    }
    {
        QuerySegmentSpec intervalSpec = new MultipleIntervalSegmentSpec(Arrays.asList(basicSchema.getDataInterval()));
        List<AggregatorFactory> queryAggs = new ArrayList<>();
        LongSumAggregatorFactory lsaf = new LongSumAggregatorFactory("sumLongSequential", "sumLongSequential");
        BoundDimFilter timeFilter = new BoundDimFilter(Column.TIME_COLUMN_NAME, "200000", "300000", false, false, null, null, StringComparators.ALPHANUMERIC);
        queryAggs.add(new FilteredAggregatorFactory(lsaf, timeFilter));
        TimeseriesQuery timeFilterQuery = Druids.newTimeseriesQueryBuilder().dataSource("blah").granularity(Granularities.ALL).intervals(intervalSpec).aggregators(queryAggs).descending(false).build();
        basicQueries.put("timeFilterAlphanumeric", timeFilterQuery);
    }
    {
        QuerySegmentSpec intervalSpec = new MultipleIntervalSegmentSpec(Arrays.asList(new Interval(200000, 300000)));
        List<AggregatorFactory> queryAggs = new ArrayList<>();
        LongSumAggregatorFactory lsaf = new LongSumAggregatorFactory("sumLongSequential", "sumLongSequential");
        queryAggs.add(lsaf);
        TimeseriesQuery timeFilterQuery = Druids.newTimeseriesQueryBuilder().dataSource("blah").granularity(Granularities.ALL).intervals(intervalSpec).aggregators(queryAggs).descending(false).build();
        basicQueries.put("timeFilterByInterval", timeFilterQuery);
    }
    SCHEMA_QUERY_MAP.put("basic", basicQueries);
}

Example 5 with BenchmarkSchemaInfo

use of io.druid.benchmark.datagen.BenchmarkSchemaInfo in project druid by druid-io.

the class SearchBenchmark method setupQueries.

private void setupQueries() {
    // queries for the basic schema
    final Map<String, SearchQueryBuilder> basicQueries = new LinkedHashMap<>();
    final BenchmarkSchemaInfo basicSchema = BenchmarkSchemas.SCHEMA_MAP.get("basic");
    final List<String> queryTypes = ImmutableList.of("A", "B", "C", "D");
    for (final String eachType : queryTypes) {
        basicQueries.put(eachType, makeQuery(eachType, basicSchema));
    }
    SCHEMA_QUERY_MAP.put("basic", basicQueries);
}